Objectives:We evaluated central venous oxygen saturation (Scvo(2)) and lactate levels as a combination measure to predict mortality in patients with severe sepsis or septic shock.Methods:We included patients older than 18 years of age who presented to a single tertiary emergency center with septic shock or severe sepsis and received early goal-directed therapy. We classified the sample into four groups according to lactate (cut-off: 4mmol/L) and Scvo(2) (cut-off: 70%) levels at the time of initial resuscitation: Group 1, high-Scvo(2), and low-lactate; Group 2, low-Scvo(2), and low-lactate; Group 3, high-Scvo(2), and high-lactate; Group 4, low-Scvo(2), and high-lactate. The primary outcome was 28-day mortality determined by multivariable Cox-regression analysis.Results:A total of 880 patients were included in this study. The 28-day mortality was 6.7% in Group 1, 15.7% in Group 2, 26.7% in Group 3, and 25.5% in Group 4 (P<0.01). Compared with Group 1, all other groups showed significant differences in mortality (P<0.01 by the log-rank test). There was, however, no difference between Groups 3 and 4. Multivariable Cox regression analysis showed that all other groups exhibited significantly increased hazard ratios for 28-day mortality, compared with Group 1.Conclusions:Oxygenation category, as represented by initial Scvo(2) and lactate levels, was significantly associated with 28-day mortality in patients with severe sepsis or septic shock. Associations between Scvo(2) 70% and 28-day survival were observed only in patients without severe lactic acidosis.